The ischaemic bed size (myocardium at risk) and collateral flow are major determinants of experimental myocardial infarct size. These parameters were compared in dogs with coronary occlusions at four commonly used sites. Regional blood flow was measured 5 min after proximal coronary occlusion of the circumflex (LCC) or anterior descending (LAD) arteries, or after distal occlusion of the LAD, or its apical branch. Occlusions were done sequentially in random order separated by 30 min intervals. The anatomic regions normally supplied by each occluded artery (ischaemic beds) were identified by simultaneous postmortem coronary perfusion with different coloured dyes. Proximal occlusions of the LCC and LAD produced equivalent areas of ischaemia (37 and 36% of the left ventricle) and these ischaemic beds had similar amounts of collateral flow (0.12 and 0.16 cm3.min-1.g-1). In comparison, distal LAD and apical branch occlusions involved 23 and 11% of the LV and collateral flow averaged 0.23 and 0.36 cm3.min-1.g-1. Thus, proximal LAD or LCC occlusion produce comparable areas of severe ischaemia. Experimental models using distal occlusions are characterised by more variable but often less severe ischaemia.